3차원 안면자동인식기(3D-AFRA)의 안면 표준점 인식 정확도 검증

Abstract

Point Recognition Precision Test of 3D Automatic Face Recognition Apparatus(3D-AFRA) Seok Jae-Hwa, Cho Kyung-Rae*, Cho Yong-Beum*, Yoo Jung-Hee, Kwak Chang-Kyu,Lee Soo-Kyung, Kho Byung-Hee, Kim Jong-Won**, Kim Kyu-Kon†, Lee Eui-Ju Dept. of Sasang Constitutional Medicine, College of Korean Medicine, Kyung Hee Univ., *TNTech, **Dept. of Sasang Constitutional Medicine, College of Korean Medicine, Dongeui Univ., †Dept. of Information Statistics, Dongeui Univ. 1. Objectives The Face is an important standard for the classification of Sasang Contitutions. Now We are developing 3D Automatic Face Recognition Apparatus to analyse the facial characteristics. This apparatus show us 3D image of man's face and measure facial figure. We should examine accuracy of position recognition in 3D Automatic Face Recognition Apparatus(3D-AFRA). 2. Methods We took a photograph of Face status with Land Mark by using 3D-AFRA. And We scanned Face status by using laser scanner(vivid 700). We analysed error average of distance between Facial Definition Points. We compare the average between using 3D-AFRA and using laser scanner. So We examined the accuracy of position recognition in 3D-AFRA at indirectly. 3. Results and Conclusions The error average of distance between Right Pupil and The Other Facial Definition Points is 0.5140mm and the error average of distance between Left Pupil and The Other Facial Definition Points is 0.5949mm in frontal image of face. The error average of distance between Left Pupil and The Other Facial Definition Points is 0.5308mm and the error average of distance between Left Tragion and The Other Facial Definition Points is 0.6529mm in laterall image of face. In conclusion, We assessed that accuracy of position recognition in 3D-AFRA is considerably good.